重症心脏瓣膜置换术后患者预后的影响因素分析及风险模型的建立

Analysis of influencing factors and establishment of risk model for prognosis of patients with severe heart valve replacement

目的 探讨并建立重症心脏瓣膜置换术(heart valve replacement,HVR)后患者预后风险模型及分析其影响因素。方法 选取北京大学深圳医院心外科2015年2月到2021年2月收治的264例重症心脏瓣膜病患者作为研究对象,调取所有患者的临床资料,通过单因素与多因素Logistic回归分析筛选出影响重症HVR患者术后死亡的危险因素,通过R软件建立相应的列线图风险预测模型,并对模型进行评估与验证。结果 Logistic回归分析结果显示,年龄(OR=4.805,95%CI:1.865~12.381)、术前心功能分级(OR=4.632,95%CI:1.887~11.373)、主动脉阻断时间(OR=1.920,95%CI:1.323~2.787)、体外循环时间(OR=1.822,95%CI:1.105~3.004)、抗凝不当(OR=3.373,95%CI:1.623~8.607)为患者术后发生死亡事件的独立风险因素(P<0.05)。基于该危险因素建立列线图预测模型,并对该模型进行验证评估。结果显示,本研究中建模集和验证集的C-index指数分别为0.957(95%CI:0.865~1.049)和0.983(95%CI:0.934~1.032),校正曲线拟合较好,两组受试者工作特征曲线(receiver operating characteristic curve,ROC)的曲线下面积(area under the curve,AUC)分别为0.957和0.983,提示本次模型具有良好的预测精准度。结论 影响重症心脏瓣膜病患者HVR术后死亡的危险因素很多,本研究基于上述危险因素建立的列线图预测模型具有一定的预测性,医护人员可根据预测模型给予高风险人群相应的预防措施,改善患者预后。

Objectives To explore and establish a prognostic risk model for patients after severe heart valve replacement(HVR)and to analyze the influencing factors. Methods Totally 264 patients with severe heart valve disease admitted in Peking University Shenzhen Hospital from February 2015 to February 2021 were selected as the research object.All the clinical data of the patients were obtained. Risk factors for mortality postoperatively in patients with severe HVR were analyzed by single factor and multiple factors Logistic regression analysis. The corresponding nomogram risk prediction model was established through R software and was evaluated and verified. Results Logistic regression analysis showed that age(OR=4.805,95%CI:1.865-12.381),preoperative cardiac function grade(OR=4.632,95%CI:1.887-11.373),aortic occlusion duration(OR=1.920,95%CI:1.323-2.787),extracorporeal circulation duration(OR=1.822,95%CI:1.105-3.004) and improper anticoagulation(OR=3.737,95%CI:1.623-8.607)were independent risk factors for postoperative death(P<0.05). Based on the risk factor,the prediction model of the histogram was established,and the model was verified and evaluated. The results showed that the C-index of the modeling set and the validation set in this study were 0.957(95%CI:0.865-1.049)and 0.983(95%CI:0.983)respectively. The area under the receiver operating characteristic curve(ROC)of the two groups was 0.957 and 0.983,respectively,indicating that the model had a good prediction accuracy. Conclusions There are many risk factors affecting the death of patients with severe valvular heart disease after HVR surgery. The prediction model of the histograph established in this study based on the above risk factors is predictive to a certain extent. Medical staff can give corresponding preventive measures to high-risk groups according to the prediction model to improve the prognosis of patients.